Increasing Numbers of Synaptic Puncta during Late-Phase LTP Ozlem Bozdagi, Weisong Shan, Hidekazu Tanaka, Deanna L. Benson, George W. Huntley  Neuron  Volume 28, Issue 1, Pages 245-259 (October 2000) DOI: 10.1016/S0896-6273(00)00100-8

Figure 1 L-LTP Is Associated with Increased Number of N-Cadherin and Synaptophysin Puncta (A) Bath application of Sp-cAMPS (50 μM; 15 min) produced a persistent potentiation of the extracellular field EPSP initial slope recorded from stratum radiatum, which reached maximum within 60–90 min (n = 9; error bars = SEM); bar indicates the duration of Sp-cAMPS application. Inset: representative EPSP traces were recorded before and 90 min after Sp-cAMPS application (arrow; calibration: 10 ms, 0.2 mV). (B) Confocal microscope images through stratum radiatum of area CA1 showing increased numbers of N-cadherin and synaptophysin-immunolabeled puncta in potentiated (L-LTP) slices in comparison with unstimulated control slices. Scale bar: upper row = 5 μm; lower row = 2.5 μm. (C) Quantitative analyses verified the number of labeled synaptic puncta in potentiated slices (n = 5) was significantly larger in comparison with unstimulated control slices (n = 12; mean ± SEM). Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 2 N-Cadherin Localizes to the Synaptic Junctional Complex in Potentiated Slices (A) Confocal images from potentiated slices showing N-cadherin (green) and synaptophysin (red) immunolabeling separately and in an overlay where regions of colocalization appear yellow. Most N-cadherin and synaptophysin puncta codistribute in a partially overlapping pattern (shown in higher magnification inset), indicating that N-cadherin localizes to the synaptic juntional complex. Scale bar = 5 μm. (B) Immunogold localization of N-cadherin from a potentiated slice demonstrating pre- and postsynaptic localization of N-cadherin at a synaptic junctional complex. Scale bar = 300 nm. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 3 Ongoing GluR-Mediated Synaptic Activity Is Required for Sp-cAMPS-Induced Increase in Synaptic Puncta Number Transiently exposing slices for 60 min to the NMDA receptor antagonist APV (50 μM) and the AMPA/kainate receptor antagonist CNQX (100 μM) inhibits L-LTP and prevents the increase in the numbers of N-cadherin puncta (left) and synaptophysin puncta (right) normally induced by exposure to Sp-cAMPS. Data are mean ± SEM. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 4 Increase In Numbers of Labeled Synaptic Puncta Requires PKA Activity and Protein Synthesis (A) The potentiation in field EPSP slope induced by bath application of Sp-cAMPS (circles) is blocked by the protein synthesis inhibitor cyclohexamide (60 μM; squares) or the protein kinase A antagonist Rp-cAMPS (100 μM; triangles). Bars indicate periods of application of Sp-cAMPS, cyclohexamide, or Rp-cAMPS (n = 4 for each group; error bars = SEM). (B and C) Inhibitors of protein synthesis and PKA that block L-LTP also block the Sp-cAMPS-induced increase in puncta numbers. Quantitative analyses demonstrate that in the presence of cyclohexamide or Rp-cAMPS, the number of N-cadherin (B) or synaptophysin (C) puncta in Sp-cAMPS-treated slices is not significantly different in comparison with that from slices treated with blockers alone or unstimulated control slices (p > 0.1) but is significantly less than the numbers observed in potentiated slices (asterisks, p < 0.01; n = 4 per condition, mean ± SEM). Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 5 L-LTP Is Accompanied by an Increase in N-Cadherin Synthesis and Dimer Formation (A) Immunoblot of unstimulated (lane 1) and Sp-cAMPS-potentiated (lane 2) slices. Protein fractions from the same samples were immunoprecipitated with an antibody against the EC1 domain of N-cadherin (lanes 3 and 4). Molecular weight markers: 205, 116, and 66 kDa. (B) Quantitative analysis of immunoblots. There is a 27% ± 5.0% increase in N-cadherin level in potentiated slices relative to unstimulated control slices (mean ± SEM, n = 6 blots, p < 0.002). (C) Metabolic labeling by 35S-methionine incorporation followed by immunoprecipitation with the same N-cadherin antibody used in (A). Three bands (127 kDa, arrowhead) increase in intensity in slices removed 15 min (lane 2) and 30 min (lane 3) following Sp-cAMPS exposure in comparison with unstimulated control slices (lane 1). See text for explanation of bands. Treatment with cyclohexamide (CHX) prior to Sp-cAMPS exposure blocks N-cadherin synthesis (lane 4). Molecular weight markers: 205 and 116 kDa. (D) Quantitative analysis of metabolic labeling experiment shown in (C). Band intensity was normalized to lane intensity. (E) Analysis of the Triton-insoluble fraction reveals the strand dimer form of N-cadherin (230 kDa, arrowhead) in both unstimulated and potentiated slices. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 6 Treatment with N-Cadherin Antibodies Prevents Sp-cAMPS-Induced L-LTP (A) Incubation of slices in one of two different N-cadherin antibodies (Ab 1258 [open squares] or Ab 1260 [open triangles]) blocks Sp-cAMPS-induced L-LTP (n = 4 for each antibody, error bars = SEM, p < 0.05, one-way ANOVA). Sp-cAMPS induced potentiation was unaffected by incubation with preimmune guinea pig sera (1258 [closed squares] or 1260 [closed triangles]) or with a control N-cadherin antibody recognizing an intracellular portion of the molecule (diamonds) (n = 4 for each group; error bars = SEM). (B) Normal synaptic properties were unaffected by antibody treatment. Input-output curves indicating the relationship between stimulus intensity and field EPSP slope (mV/ms) show that slices incubated in Ab 1258 (closed squares) were not significantly different from normal bath-maintained control slices (open squares) (n = 4). (C) Representative Western blot showing specificity of antibody 1260. Blot shows a single band of ∼127 kDa (arrow) that is the approximate size of N-cadherin. Immunostaining could be competed out with purified N-cadherin EC1 protein but not with BSA. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 7 N-Cadherin Antibodies Significantly Attenuate Adhesive Aggregation of Mouse L Cells Stably Transfected with N-Cadherin In the presence of Ca2+ and Ab 1260, the amount of aggregation is not significantly different than that of transfected L cells without Ca2+ but is significantly attenuated in comparison with transfected L cells incubated in Ca2+ alone or Ca2+ plus preimmune serum (one-way ANOVA). Data are mean ± SEM. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 8 N-Cadherin Antibodies Attenuate L-LTP Induced by Electrical Stimulation without Affecting E-LTP Tetanic stimulation (double arrows) was used to produce both E-LTP and L-LTP. Function-blocking antibodies did not block E-LTP but significantly attenuated potentiation ∼60 min posttetanization (L-LTP), contemporaneous with the transition from E-LTP to L-LTP. In contrast, slices incubated in preimmune serum (closed squares) exhibited both E- and L-LTP; (n = 4; error bars = SEM, p < 0.05, one-way ANOVA). Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 9 Speculative Mechanisms Leading to Increasing Numbers of Synaptic Puncta during L-LTP Two possible models (A and B) to account for the increased puncta number during L-LTP. In both models, the membrane-permeable Sp-cAMPS activates an intracellular, cAMP-dependent signaling cascade that leads to a rise in N-cadherin protein levels. In the first model (A), N-cadherin and synaptophysin are recruited to new synapses that form de novo. Alternatively (B), some preexisting synapses may be subthreshold for detection in the unstimulated state (left), possibily because the synapses are too small and contain too few molecules to be detected by the markers used. However, during L-LTP these synapses become detectable through molecular remodeling involving an accumulation of N-cadherin and synaptophysin (right). Regardless of which mechanism (A or B) leads to L-LTP, potentiation is blocked in the presence of the function-blocking N-cadherin antibodies (C), possibly because the antibodies prevent the formation of the adhesive interface between pre- and postsynaptically apposed cadherin molecules, thus interfering with the appropriate spacing of the synaptic cleft or altering the area of the active zone (question mark). The increase in numbers of labeled synaptic puncta is still evident in L-LTP-blocked slices, presumably because Sp-cAMPS still activates its intracellular signal cascade both pre- and postsynaptically. For simplicity, the intracellular binding partners of the cadherins are not depicted. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)

Figure 10 Thresholding Analysis of Immunoreactive Puncta Single optical section of a control slice (A and C) or an Sp-cAMPS-potentiated slice (B and D) immunoreacted with N-cadherin and visualized with Oregon green. A sample series of images from different slices was used to establish optimal brightness and contrast levels. The images shown in (A) and (C) were imported into NIH Image; the density slice function was used to highlight individual puncta (shown in red in [B] and [D]). This was achieved by establishing a threshold that yielded the greatest number of individual puncta without causing a fusion of puncta. Neuron 2000 28, 245-259DOI: (10.1016/S0896-6273(00)00100-8)